Creative is the new targeting. The settings stopped mattering.

01What the phrase really means

This is not about prettier ads. It is how the machine works now. Your ad opens with, if you are a video editor using Premiere Pro. You just targeted video editors who use Premiere Pro. Not with a settings panel. With 1 sentence. Who stops tells the platform much more than who scrolls past. Inside a day or 2 it knows who this ad is for better than any audience you could build.

The sentence did the targeting. The settings panel only watched.

2 things happened together. The system got much better at guessing who will act, and a hand built audience stopped telling it anything new. Worse, every box you tick removes people it would have found.

Now the trap, the reason people try this and call it nonsense. Broad targeting and cohort specific ads are 1 move, not 2. Go broad while you still run 2 or 3 general ads and the account gets worse, because you removed the only steering you had. Go broad with 10 or 15 ads that each name a specific kind of person and the steering comes back, now in the words.

02The settings job turned into a research job

The work moved out of the ads manager into 3 questions, and none of them can be answered inside the account.

Where do the answers live? Sales calls. Client group chats. Review sites. The comments under your posts. Almost never a keyword tool. I keep those lines in our company brain inside the AiOS command center, and a Telegram assistant hands them back when I write. The listening is the point, not the tool.

03A cohort is not a demographic

People read cohort and hear age, gender and city. A cohort is a person, at a certain awareness level, at a certain moment. All 3 parts matter.

Someone seeing your ad at 8 in the evening at dinner with family is not the same person 1 hour later. How many times they bumped into you changes them again. 3 touchpoints in, they are someone else. So 1 human can sit in 3 cohorts. Age and city cannot say that. Awareness can.

The usual mistake is to think the viewer is closer to buying than they are. A microphone ad that says look at the sound quality, buy this, only lands on someone who already knows the brand and the problem. Almost nobody is that person.

The ad is not bad. It is pitched at the wrong level.

Real cohorts come from listening. On a client group chat, someone said in passing that their first step out of bed hurts. That became an ad angle the same day. Nobody in a group chat is performing. People describe their problem the way they would to a friend, which is how a good ad should sound.

Then keep that specificity to the end. Back pain is a category. At 42, getting better stops being about rest and starts being about which muscle stopped doing its job, that is a cohort. The longer line speaks to fewer people and wins, because the person it describes sees themselves in it.

04The 5 things on paper before you write a line

If the ad carries the targeting, the research carries the ad. When ads are not converting, the cause is almost never the account. Nobody did the research the ad needed.

I know how basic that list looks. But this is exactly what is missing. Everyone jumps to hooks and edits and thumbnails, and nobody did the work that tells them what to say.

1 more thing sits under all 5. Babies are the consumers of a diaper. Parents are the customers. Your ad has to know which of the 2 it is talking to, because the pain, the want and the benefit are different.

05Volume is reach, not extra work

If the ad is the targeting, the number of ads is your reach. What feels like a production headache is really coverage. And 15 ads is not 15 times the work if you build them in parts.

There is a free version too. Post the 15 cohort lines as 15 slides of 1 carousel. A carousel breaks its numbers down slide by slide, so 15 cohorts becomes 15 tests, run for free, in front of people who already know you. The first slide is always inflated, because everybody sees it. Look at the spikes after it, where somebody had to swipe to arrive.

This is the structure we run on our own account. You record 1 body once, the part that explains what you do and why it works. Then you swap only the first 5 to 7 seconds for each cohort, and that opening is a real story about somebody in that cohort. 1 might be a call with the co-founder of an insurance provider who wanted a CFO dashboard. Another, a managing director in Dubai who needed renewals automated. Same body, different first 7 seconds. The story has to be real and from this week. Make it up and it sounds like every other ad.

A plain text ad wins when the words nail the cohort. A heavily edited emotional video can lose to it, and we have watched it in our own accounts. Polish says, this is an ad. Being specific says, this is about me. In a feed the second one wins.

06Squeeze the winner, then prove it

When something works, the instinct is to make the next new thing. That is expensive and mostly wrong. Most of your effort should go into reskinning what already won. Cut the first 3 seconds off a winner and paste it onto everything else. Rebuild 1 strong piece of proof many times: recoloured, new headline, new font. Split time and money 70% on the proven thing, 20% on close copies, 10% on new ideas. Do the 70 first.

Give a test about 10 days before you judge it. Then read cost per result by angle, not by ad. For a local service business that is cost per appointment. Then the part most people skip. Hand the winning angle to your organic side, the work I call STORM. Organic shows cheaply which versions have life, and the winners come back into ads already proven.

1 thing quietly breaks that loop. If all 15 ads speak to people ready to buy today, you finish that small pool and results fall off a cliff. It looks like ad fatigue when it is not.

Now the limit. The real wall when you spend more is trust. The further you push, the colder the people you reach. Your promise is not special either. Everybody promises the same outcome in the same words. So add proof, not a cleverer promise. In person beats virtual, live beats recorded, someone who looks like your buyer beats someone who does not, and the exact outcome they want beats one that is close. It matters more now that people ask a chatbot who to hire before they ask a friend.

Being specific gets you attention. Proof gets you belief.

I have taught broad targeting plus creative volume for 3 years, and it is only going mainstream now. The ideas that work are around long before they are popular. So get good at listening. Keep a document of the exact phrases prospects use. Not summaries. Phrases. And promote the person who understands your customers, not the one who knows where the buttons are. Button knowledge lasts months. Customer knowledge pays for years.

07Questions, answered